Effects of Backbone Contacts 3′ to the Abasic Site on the Cleavage and the ProductBinding by Human Apurinic/Apyrimidinic Endonuclease (APE1)†

Tadahide Izumi,*,# Catherine H. Schein,§ Numan Oezguen,§ Yanling Feng,# and Werner Braun§

Sealy Center for Molecular Science, Sealy Center for Structural Biology, and Department of Human Biological Chemistry &Genetics, UniVersity of Texas Medical Branch, GalVeston, Texas 77555-1079

ReceiVed April 22, 2003; ReVised Manuscript ReceiVed NoVember 19, 2003

ABSTRACT: The mammalian apurinic/apyrimidinic (AP) endonuclease (APE1) is a multifunctional proteinthat plays essential roles in DNA repair and gene regulation. We decomposed the APEs into 12 blocks ofhighly conserved sequence and structure (molegos). This analysis suggested that residues in molegoscommon to all APEs, but not to the less specific nuclease, DNase I, would dictate enhanced binding todamaged DNA. To test this hypothesis, alanine was substituted for N226 and N229, which form hydrogenbonds to the DNA backbone 3′ of the AP sites in crystal structures of the APE1/DNA complex. Whilethe cleavage rate at AP sites of both N226A and N229A mutants increased, their ability to bind to damagedDNA decreased. The ability of a double mutant (N226A/N229A) to bind damaged DNA was furtherdecreased, while theVmax was almost identical to that of the wild-type APE1. A double mutant at N226and R177, a residue that binds to the same phosphate as N229, had a significantly decreased activity andsubstrate binding. As the affinity for product DNA was decreased in all the mutants, the enhanced reactionrate of the single mutants could be due to alleviation of product inhibition of the enzyme. We concludethat hydrogen bonds to phosphate groups 3′ to the cleavage site is essential for APE1’s binding to theproduct DNA, which may be necessary for efficient functioning of the base excision repair pathway. Theresults indicate that the molego analysis can aid in the redesign of proteins with altered binding affinityand activity.

DNA in all living organisms is continuously attacked byreactive oxygen species (ROS)1 and environmental geno-toxicants such as DNA alkylating reagents (1-3). Thesereagents generate abnormal bases, apurinic/apyrimidinic (AP)sites, and DNA strand breaks, which are either toxic ormutagenic. To maintain their genomic integrity, all organismshave a similar base excision repair (BER) pathway to removedamaged sites in DNA (4, 5). In the first step of thisstreamlined pathway (6, 7), DNA glycosylases cleave theN-glycosylic bonds of abnormal bases to generate either APsites or DNA strand breaks byâ- or âδ-elimination reaction(8, 9). Next, AP endonucleases (APEs) generate 3′-OHtermini that serve as primers for subsequent gap fillingreactions by DNA polymerases (7), followed by strand-sealing reactions by DNA ligases (10).

APE1, a key player in the mammalian BER pathway, isan essential protein; embryos of APE1 homozygous knockoutmice died during early embryonic development (E3.5-E5.5)

(11-13). In addition to its role in DNA repair, APE1 alsofunctions as a redox gene regulator to activate such criticaltranscriptional factors as cJun/cFos, p53, and HIF1-R (14-16). APE1 is also a transcriptional repressor of the parathy-roid hormone gene and possibly of the APE1 gene itself (17-19). The redox mechanism, which apparently requires a Cysresidue in mammalian APE1, has not been completelyestablished (20). The mechanism of the repressor functionhas been analyzed and reported to involve acetylation ofN-terminal Lys residues (21).

The residues of APE1 that are essential for cleavage ofDNA at AP sites have been identified by biochemical andstructural studies (22-28), in particular by crystal structuresof APE1 bound to substrate DNA (28). Most of the residuesknown to be required for hydrolysis and Mg2+ coordination,such as E96, D210, and H309 were easily identified by theirconservation in the sequences of APE family members,includingEscherichia coliexonuclease III (Xth). In additionto conserved residues, a structural study revealed that Arg177 (R177) penetrates the substrate DNA through the majorgroove, and forms a hydrogen bond to the DNA backbone(28). The residue is not required for the AP endonucleaseactivity, but crucial for product DNA binding (28), which isbelieved to be a common feature among many BER enzymesfor a coordinated repair process. It is possible that otherresidues are also required for APE1’s high affinity for thecleaved DNA. However, simple sequence conservationanalysis does not reveal how APE1 specifically recognizesthe substrate and cleaved DNA.

† This work was supported by DOE Grant (DE-FG-03-00ER63041)and NIH Grant CA53791 and CA98664.

* Corresponding author: Tadahide Izumi, Sealy Center for MolecularScience, University of Texas Medical Branch, Galveston TX 77555-1079; phone: (409) 772-6308; fax: (409) 747-8608; E-mail: taizumi@utmb.edu.

# Sealy Center for Molecular Science.§ Sealy Center for Structural Biology.1 Abbreviations: AP site, apurinic/apyrimidinic site; APE, AP

endonuclease; BER, base excision repair; CD, circular dichroism;E.coli, Escherichia coli; EMSA, electrophoresis gel retardation assay;ROS, reactive oxygen species; Xth, exonuclease III.

684 Biochemistry2004,43, 684-689

10.1021/bi0346190 CCC: $27.50 © 2004 American Chemical SocietyPublished on Web 12/24/2003

We used a different approach to the problem, whichinvolved identifying discrete conserved areas in a sequencealignment of many APEs from diverse sources with our web-based MASIA tool (29). Ten of the 12 conserved motifs soidentified were then shown to be molecular building blocks,or “molegos”, as they formed distinct secondary structuralelements in the crystal structures of Xth and APE1 aloneand in complex with the substrate DNA (29). Five molegosare common to DNase I (29). Comparison of the crystalstructures of complexes of human APE1 and DNase I withtheir DNA substrates indicated that these molegos wereessential components for the nucleolytic activity of bothproteins. Indeed, some of the molegos are shared by distantlyrelated enzymes such as inositol-5′-polyphosphate phos-phatases (29), that catalyze phosphorolytic cleavage, but havedifferent substrate specificity than APE1. This modularapproach, in which blocks of functional units (molegos) arefirst identified and then evaluated in detail, is an efficientmethod to determine the molecular basis of functionaldiversity among similar enzymes. By comparing APE1’s sidechain composition to that of the less specific endonuclease,DNase I, we were able to identify a region of APE1 thatcontributes to product binding specifically. The present studyreports that mutations in these regions in APE1 affect bothsubstrate and product binding.

EXPERIMENTAL PROCEDURES

Construction of Mutants by PCR.The source of the wild-type (WT) and R177A human APE1 cDNA was describedearlier (30). Other missense mutants, including N226A,N229A, N226A/N229A, and R177A/N226A, were createdby PCR-cloning with specific primers. Briefly, a vectorprimer and an oligonucleotide primer (5′ CCC CTT AGGGTT GCG AAG GTC AAT 3′), which anneals just upstreamof the N226 or N229 site of the hAPE1 cDNA to introducea Bsu36I site, was used for PCR-cloning of the upstreampart of the hAPE1 cDNA using a pTrcHis2-TOPO cloningkit (Invitrogen). Then primers for N226A (5′ CGC AACCCT AAG GGG GCG AAA AAG AAT GCT GGC TTC3′), N229A (5′ CGC AAC CCT AAG GGG AAC AAAAAG GCG GCT GGC TTC 3′), and N226A/N229A (5′ CGCAAC CCT AAG GGG GCG AAA AAG GCG GCT GGCTTC 3′) were used to amplify the downstream part of thehAPE1 cDNA with a vector primer, and cloned into thepTrcHis2 derivative usingBsu36I. The DNA containingmissense mutations were finally cloned into the pET15bprotein expression vector (Novagen) as described previously(22). The DNA sequences of the mutant cDNA wereconfirmed by UTMB’s Protein Chemistry Core laboratory.

Purification of Proteins.APE1 proteins were expressedand purified by the method described earlier (22) with slightmodifications. Briefly,E. coli BL21/codon plus was trans-formed with the pET15b-derivatives, and the bacteria cultures(500 mL in LB broth) were grown to OD600 of 0.3-0.5.Then the proteins were induced with 0.5 mM IPTG at 16°C for 14-16 h (31, 32). Cells were suspended in asuspension buffer containing 20 mM Tris (pH 8.0) and 1 MNaCl, and broken by sonication. After centrifugation, theextracts were applied to a 3-mL column of Ni-NTA. Thecolumn was washed with 18 mL of the suspension bufferand with the same buffer cotaining 20 mM imidazole. TheAPE1 protein was then eluted with 10 mL of the suspension

buffer with 200 mM imidazole, and then dialyzed against20 mM Tris-Cl (pH 8.0), 150 mM NaCl, and 1 mMdithiotreitol (DTT). The histidine tag peptides were removedby thrombin at 4°C for overnight incubation, and thefractions were further purified to homogeneity with SP-Sepharose column chromatography using the AKTA purifiersystem (Pharmacia). The final fractions were dialyzed against20 mM Tris (pH 8.0), 300 mM NaCl, 1 mM DTT, and 50%glycerol, and stored at-20 °C.

Circular Dichroism (CD) Spectrum.The secondary struc-ture of the APE1 proteins were monitored with an AVIV 62DS circular dichroism spectropolarimeter. The far-UV CDspectra (195-260 nm) were obtained in fused quartz cuvetteswith 0.1-cm path length with protein solutions in 20 mMTris-Cl (8.0) and 150 mM NaCl. Each spectrum was recordedwith a 1-nm increment. For each sample (with a concentra-tion of about 200µg/mL), three repetitive scans wereobtained and averaged. At least two independent experimentswere carried out.

AP-Endonuclease ActiVity. For kinetics study, a 43-meroligonucleotide containing the AP site analogue, tetrahydro-furan, was used as described previously (22, 33). The oligowas 5′-end-labeled with [γ-32P]ATP and purified by gelfil-tration using Sephadex G25. The substrate for the AP-endonuclease assay was prepared by annealing the 43-meroligonucleotide (Midland Corp) with the reverse comple-mentary oligonucleotide (22), followed by purification usingnondenaturing polyacrylamide gel electrohoresis (PAGE).APE1 proteins (22 pM) were mixed with the substrates withconcentration (indicated in Figure 3 as [S]), and incubatedat 37°C for 0-4 min. The reaction within this time periodexhibited a linear increase of the product DNA (data notshown). The reactions were stopped by the addition of thestop buffer (10 mM EDTA in 88% formamide solution with0.05% bromophenol blue) and kept on ice until the sampleswere analyzed in 8% PAGE containing 8 M urea to separatethe substrate DNA from the cleaved product. The gels weredried and then analyzed with a phosphorescence detectionsystem (Storm, Molecular Dynamics). The kinetics data werefitted by nonlinear least-squares regression to obtainVmax

andKm using the Michaelis-Menten equation.Electorophoretic Mobility Shift Assay (EMSA).The 5′-

[32P]-labeled 43-mer substrate DNA (about 25 pM) wasincubated with purified WT-APE1 and the indicated mutantsthereof (0-500 nM as shown in Figure 4) in the bindingbuffer containing 25 mM HEPES, 50 mM NaCl, 4% Ficoll,1 mM DTT, 2.5 mM EDTA for 30 min at 25°C. Thesamples were then run in nondenaturing 5% PAGE in thebuffer containing 6.7 mM Tris, 3.3 mM sodium acetate, 1mM EDTA at 4 °C. The gels were dried and then analyzedwith the Storm system. The data were fittedusing Sigmaplotin the equation:

where f equals to the ratio of bound DNA to total DNA([DNA] B/[DNA] T), [E] is the concentration of the protein,andKd is the dissociation constant. The protein concentrationwas in excess of that of the bound DNA. APE1’s binding tothe product DNA (after the hydrolysis reaction) was observedwith the AP site-containing DNA (either the 43-mer oligo-

f )[E]

[E] + Kd

DNA Binding Specificity of Human APE1 Biochemistry, Vol. 43, No. 3, 2004685

nucleotide or 5′-Cy5 labeled 26-mer, IDT technology) in thesame buffer containing 2 mM MgCl2 and no EDTA. Thesequence of the 26-mer DNA is 5′ TCCTCAGTTTCACTX-CTGCACCGCAT, where X denotes tetrahydrofuran.

RESULTS

N226 and N229 Form Hydrogen Bonds to DNA BackboneDownstream to AP Sites.Of the molegos specific to APE1,two molegos were in contact with the two bases 3′ to theAP site on the damaged strand in the cocrystal structure ofthe protein with its DNA (29). One residue in molego 7,N226, formed hydrogen bonds to the phosphate backbonein three cocrystal structures, while N229, from a neighboring,less conserved sequence area, was within hydrogen bondingdistance in only one (1DE8, Figure 1). Although an exactsequence alignment between APE1 and Xth in this area isnot possible, N226 probably corresponds to N167 in thelatter. In the 1DE8 crystal structure of hAPE1 bound to thesubstrate DNA, N229 binds to the same phosphate, one basedownstream from the AP site, as R177, mutation of whichin previous studies (28) was shown to increaseVmax butdecrease binding to its substrate.

N226 and N229 Mutants Showed Higher Vmax and Km

Compared to the WT Protein.To examine the importanceof these contacts in APE1 catalysis, we constructed alaninesubstitution mutants, namely, N226A, N229A, and the doublemutant N226A/N229A. These mutants and the WT proteinwere soluble when expressed inE. coli and could be purifiedto homogeneity by the same protocol (Figure 2). Both Asnresidues are on the surface of the protein, so we expectedno major conformational change. This was confirmed bycircular dichroism (CD) spectra (Figure 2B). The spectra forthe mutant proteins were essentially identical to that of theWT, confirming that their effect was only local and did notaffect the global fold of the protein. The spectra containeda broad minimum centered at 210 nm, consistent with theprimarily â-sheet structure of APE1. There was also ashoulder at 225-230 nm spectra for all APE1 proteins, whichcould reflect optical activity from exposed aromatic sidechains (34).

Enhanced APE ActiVity of the N226A and N229A Mutants.Because of their direct interaction with both substrate andproduct DNA in the cocrystal structures (28), we inferred

that the N226A and N229A mutations would affect APE1’saffinity for the damage-containing DNA. We thus examinedtheir AP endonuclease activity compared with the WTprotein, to probe their influence on the DNA binding. Todetermine the kinetics parameters of the APE1 proteins forthe AP endonuclease activity, we used tetrahydrofuran (THF)(22, 28, 30) because of its stability. The intact AP site iseasily cleaved spontaneously viaâ-elimination reaction, andthus is not suitable for accurate quantitation of the cleavagereaction. A previous report also showed that APE1 is ableto cleave THF with a similar efficiency to the intact AP sites(35). A 5′-end labeled oligonucleotide (43-mer) containingTHF was incubated with the purified proteins, and the APendonuclease reactions were assayed after denaturing gelelectrophoresis which separated the cleaved DNA from theuncut substrate DNA (Figure 3, Table 1). The N226A,N229A, and N226A/N229A double mutant showed higherspecific activity than the WT protein. TheVmax of the singlemissense mutant proteins were more than 2-fold higher thanthe WT, and the double mutant also showed a slightly higherVmax (Table 1). However, all three mutants had significantlyhigher Km than the WT protein. While theKm was 4- and

FIGURE 1: The hydrogen bonding network 3′ to the AP site ofdamaged DNA formed by residues N226, N229, and R177 of APE1.The contacts (and their length in angstroms) shown with thestructure of the damaged DNA are taken from the 1DE8.pdbcoordinate file (28). Colors: N226/N229 red, R177 blue, abasicsite turquoise, DNA green, hydrogen involved in H-bonds gray,and the bond to be cleaved yellow.

FIGURE 2: Purification of the WT and mutant APE1 proteins. (A)12% SDS/PAGE of the purified proteins stained with Coommassieblue. Protein amount: 2µg of the WT (W), N226A, N229A, andN226A/N229A; 1µg of R177A, and R177A/N226A (B) CD spectraof APE1 proteins including the WT, R177A, N226A, N229A,N226A/N229A, R177A/N226A.

686 Biochemistry, Vol. 43, No. 3, 2004 Izumi et al.

5.4-fold higher for the N226A and N229A mutants, respec-tively, theKm of the N226A/N229A mutant was more than8-fold higher, indicating that binding to the substrate wasprogressively weakened by removing residues that couldform the hydrogen bond network. These data are similar toresults obtained for the R177A mutant, which also showeda higherVmax /Km values than that of the WT protein (28).Therefore, we tested whether the loss of the Arg residuewould further affect the activity. The double mutant R177A/N226A, which was completely soluble and thus purified asthe single missense mutant proteins, showed a drasticdecrease in its specific activity (9× 10-2 s-1), 1/16 of theWT APE1.

Recognition of the Substrate and Product DNA by theAPE1 Proteins.All the mutants were then tested for theirability to bind substrate DNA in the electrophoretic mobilityshift assay (EMSA). The 43-mer THF-containing oligonu-cleotide was incubated with the WT and the mutant APE1proteins. The binding reaction contained 2 mM EDTA andno divalent cation (Mg2+), an essential cofactor for the APEactivity, so that the APE1 binding occurred to the substrateDNA (36). All the mutants had lower affinity for thedamaged DNA substrate than the wild type, consistent withtheir increasedKm values in the first experiments (Figure 4,Table 2). The affinity for the substrate DNA of the doublemutant, R177A/N226A, was drastically decreased (Figure4, Table 2). These data suggest that all three residues, R177,N226, and N229 contribute to proper substrate binding byAPE1, and underscore the importance of the H-bondingnetwork formed by these residues.

Binding of Mutant and wt APE1 Protein to CleaVedProduct DNA.The fact that APE1 formed a stable complexwith a cleaved DNA and Mn2+ in the crystallographic study(28) suggests that APE1 may still interact with the DNAafter its reaction and thus is not released efficiently fromthe DNA. Others have suggested that such binding may benecessary to ensure that the product is efficiently passed tothe subsequent enzymes in the base excision repair pathway(28, 37, 38). Reduced product binding could explain thehigher specific activity of the single missense mutants, asweakening APE1’s affinity for the product DNA couldenhance turnover. Thus, we carried out gel retardation assaysfor APE1/product DNA complexes to probe the importanceof the residues for the product binding. An oligonucleotidecontaining an AP site was mixed with the purified APE1proteins in the presence of 2 mM MgCl2 and no EDTA, acondition in which APE1 is fully active. Specific binding ofproduct of the mutants and WT APE1 was followed by gelelectrophoresis (Figure 5). All the mutant APE1 proteins haddecreased affinity for the cleaved DNA, and so released theirproducts after cleavage (Figure 5). Denaturing gel electro-phoresis which separates the substrate DNA and the cleavedDNA confirmed that the substrate DNA was completelycleaved by all APE1 proteins, confirming that the complexin this condition was formed between APE1 and the cleavedDNA (Figure 5).

DISCUSSION

Point Mutations Establish the Importance of the HydrogenBond Network 3′ to the CleaVage Site for Specificity.Inprevious work, we compared the contacts that APE1 madeto DNA to those made by a less specific nuclease, DNase I(29). Here we demonstrate that the contacts we identified asspecific to APE1, a network of hydrogen bonds to the

FIGURE 3: Effect of N226 and N229 on the APE activity. APendonuclease activity for the WT APE1 (A), N226A (B), N229A(C), and N226A/N229A (D) mutants. The 5′-32P-labeled 43-meroligonucleotide was incubated with reaction conditions describedin Experimental Procedures. [S]: substrate concentration (nM);V0/[E]: cleaved products released by APE1 (s-1).

Table 1: Kinetics of APE1 Proteins for AP-Endonuclease Activitya

Km (nM) Vmax (s-1)

WT 3.4( 0.48 1.38( 0.05N226A 13.6( 1.21 5.26( 0.18N229A 18.4( 3.4 2.28( 0.16N226A/N229A 27.8( 10.0 1.65( 0.18

a 5′-32P-labeled 43-mer oligonucleotide was used in the standardenzyme assay described in Experimental Procedures.

FIGURE 4: Effects of the missense mutations on substrate DNAbinding. Electrophoresis mobility shift assay carried out with theTHF containing 43-mer and APE1 proteins. Ratio of bound([DNA] B) to total DNA ([DNA]T) was plotted as a function ofenzyme concentration (indicated as log [APE1]).

Table 2: Affinity of APE1 Proteins for DNA with an AP Sitea

protein Kd-app (nM) ( STD

WT 0.31( 0.07N226A 0.55( 0.12N229A 1.33( 0.38N226A/N229A 2.20( 0.87R177A 2.15( 0.78R177A/N226A 31( 16.5

a Values calculated based on Figure 4 using Sigmaplot.

DNA Binding Specificity of Human APE1 Biochemistry, Vol. 43, No. 3, 2004687

backbone of the DNA 3′ to the AP site, do indeed contributeto the affinity of the enzyme for both substrate and product(28, 37). We created point mutations of mammalian APE1that, based on a modular decomposition analysis of thesequences of APE family members (29), should disruptbinding to the phosphate groups 3′ of the AP site in damagedDNA. We showed that single mutations, which affect theinterface between APE1 and the DNA backbone downstreamof the AP site, increased the rate of cleavage of AP sites,which was probably due to enhanced turnover of the enzyme,compared to the WT protein. A double mutant that removedbinding to two different 3′ phosphate groups greatly reducedactivity, concomitant with a drastic drop in the ability ofthe mutant to bind either substrate or product DNA.

The role of R177, which penetrates the major groove ofthe AP site containing DNA in the cocrystal structure andforms a hydrogen bond to the DNA backbone one nucleotidedownstream of the AP site (28, 38), was previously examinedfor the overall reaction rate (28). It was concluded that theresidue did not play a significant role in catalysis, becausealanine substitution for the residue did not inactivate theprotein, but enhanced its specific activity for the AP site-cleavage. The decomposition analysis of DNase I familymembers (29) revealed that in addition to R177, N226, andpossibly N229 formed a network of H-bonds 3′ to thecleavage site on the damaged DNA. No such binding wasseen in cocrystal structures of the nonspecific DNase I. Thisresult led to the hypothesis that the network was essentialfor the specific cleavage of AP-sites by APE1. Further, wesuspected that this network was also needed to enhanceproduct binding after the cleavage reaction, which was seenin a cocrystal structure of APE1 (28, 37).

As previously observed, R177A mutant had reducedaffinity (higher Km) for substrate DNA, while theVmax wasenhanced. We hypothesized that release from the productinhibition must account for the enhanced rate of cleavageobserved for the mutant. In this work, a double mutant ofR177 with another residue with similar binding, N226,enabled us to determine the role of R177A in the substraterecognition more precisely. In one of the three cocrystalstructure (1DE8), N229 interacts with the same phosphate

as R177, and N226 interacts with the next phosphatedownstream in other cocrystal structures (1DE9 and 1DEW).Both N226A and N229A had higherVmax values than theWT protein and also showed higherKm values. Combiningthe N226A mutation with N229A greatly increased theKm

value for the APE reaction, and the ability to bind thedamaged DNA was decreased drastically in the N226A/R177A double mutant (Figure 3). Furthermore, the specificactivity of R177A/N226A double mutant was considerablydecreased by almost 20-fold. These results show the additiveeffects of loss of the H-bonds, and demonstrate conclusivelythat bonds 3′ of the AP-site are necessary for specific DNAbinding by APE1.

We can compare these results to those obtained for N212(39), which is absolutely conserved in all DNase I familymembers regardless of substrate specificity. Alanine substitu-tion of N212 completely inactivates the enzyme (39). Further,N212 makes a hydrogen bond to the scissile phosphate beforethe reaction that is not seen in crystal structures of the cleavedcomplex. Thus, while N212 is definitely necessary for theAP endonuclease activity of the enzyme, it does notcontribute to the unique aspects of APE1 binding to damagedDNA and the cleaved product.

Sequence Decomposition ReVeals the Important Segmentsfor Specific Binding and ActiVity. The N226/N229 side chainswere selected for point mutations based on a sequencedecomposition of multiple sequence alignments of the APEfamily and subfamily coupled with a structural analysis (29).In that study, it was shown that other enzymatic families,with a common metal-based phosphorolytic cleavage butdiffering substrate choice, share five of the molegos. Forexample, in a complex of the bacterial inositol polyphosphate5′-phosphatase (IPP) with inositol polyphosphate (IP), thegeometry of the catalytic core residues and the substrate isremarkably similar to that seen in the APE1/DNA complex.The residues in IPP near the P-O bond cleavage site ofinositol phosphate are very similar to those in APE near thescissile P-O of the AP-site (28, 40). By eliminating thesecommon areas that form a general metalophosphatase site,we could isolate those areas that are involved in generalsubstrate binding. By further eliminating areas that boundundamaged DNA in the nonspecific nuclease DNase I, wecould isolate the DNA binding residues that could mostaccount for specific binding to damaged DNA.

The methodology used here provides a practical approachto design in vitro experiments to study the amino acid basisof specific protein functions. In the present study, we haveshown that our approach is useful to produce variants of theAPE1 with altered binding specificity and activity. Theapproach is not limited to missense mutations, but allowsrigorous alteration of large segments (i.e., region replace-ment). In future work, we should be able to use this form ofanalysis to isolate areas that may be involved in the lessstudied aspects of APE activity, such as the redox based generegulation.

ACKNOWLEDGMENT

We would like to thank Dr. J. C. Lee and L. Lee for theirguidance on the CD analysis, and Dr. S. Mitra for hisinsightful discussion.

FIGURE 5: Cleaved DNA-binding by APE1 proteins. (A) 5′-Cy5labeled 26-mer oligo containing a tetrahydrofuran (260 nM) wasincubated with (1) no protein, or 20 ng of (2) the WT, (3) N226A,(4) N229A, (5) N226A/N229A in the presence of 2 mM EDTAand no divalent cation, and EMSA was carried out in nondenaturinggel electorophoresis. (B) 5′-32P-labeled 43-mer oligo with atetrahydrofuran was incubated with (1) no protein, or 20 ng of (2)the WT, (3) N226A, (4) N229A, (5) R177A, (6) R177A/N226A,and (7) N226A/N229A. (C) the samples in panel B were run indenaturing gel electrophoresis to separate the substrate and productDNA.

688 Biochemistry, Vol. 43, No. 3, 2004 Izumi et al.

REFERENCES

1. Ward, J. F. (1994)Int. J. Radical Biol. 66, 427-432.2. Breen, A. P., and Murphy, J. A. (1995)Free Radical Biol. Med.

18, 1033-1077.3. Roy, R., Biswas, T., Hazra, T. K., Roy, G., Grabowski, D. T.,

Izumi, T., Srinivasan, G., and Mitra, S. (1998)Biochemistry 37,580-589.

4. Friedberg, E. C., Walker, G. C., and Siede, W. (1995)DNA Repairand Mutagenesis, ASM Press, Washington, DC.

5. Mitra, S., Izumi, T., Boldogh, I., Bhakat, K., Hill, J., and Hazra,T. (2002)Free Radical Biol. Med. 33, 15-28.

6. Mitra, S., Boldogh, I., Izumi, T., and Hazra, T. K. (2001)EnViron.Mol. Mutagen 38, 180-190.

7. Wilson, S. H. (1998)Mutat. Res. 407, 203-215.8. Krokan, H. E., Standal, R., and Slupphaug, G. (1997)Biochem.

J. 325, 1-16.9. Hazra, T. K., Izumi, T., Maidt, L., Floyd, R. A., and Mitra, S.

(1998)Nucleic Acids Res. 26, 5116-5122.10. Tomkinson, A., and Mackey, Z. (1998)Mutat. Res. 407, 1-9.11. Xanthoudakis, S., Smeyne, R. J., Wallace, J. D., and Curran, T.

(1996)Proc. Natl. Acad. Sci. U.S.A. 93, 8919-8923.12. Ludwig, D. L., MacInnes, M. A., Takiguchi, Y., Purtymun, P. E.,

Henrie, M., Flannery, M., Meneses, J., Pedersen, R. A., and Chen,D. J. (1998)Mutat. Res. 409, 17-29.

13. Meira, L., Devaraj, S., Kisby, G., Burns, D., Daniel, R., Hammer,R., Grundy, S., Jialal, I., and Friedberg, E. (2001)Cancer Res.61, 5552-5557.

14. Xanthoudakis, S., Miao, G., Wang, F., Pan, Y. C., and Curran, T.(1992)EMBO J. 11, 3323-3335.

15. Jayaraman, L., Murthy, K. G., Zhu, C., Curran, T., Xanthoudakis,S., and Prives, C. (1997)Genes DeV. 11, 558-570.

16. Ema, M., Hirota, K., Mimura, J., Abe, H., Yodoi, J., Sogawa, K.,Poellinger, L., and Fujii-Kuriyama, Y. (1999)EMBO J. 18, 1905-1914.

17. Okazaki, T., Chung, U., Nishishita, T., Ebisu, S., Usuda, S.,Mishiro, S., Xanthoudakis, S., Igarashi, T., and Ogata, E. (1994)J. Biol. Chem. 269, 27855-27862.

18. Izumi, T., Henner, W. D., and Mitra, S. (1996)Biochemistry 35,14679-14683.

19. Kunninger, D., Izumi, T., Papaconstantinou, J., and Mitra, S.(2002)Nucleic Acids Res. 30, 823-829.

20. Ordway, J. M., Eberhart, D., Curran, T. (2003)Mol. Cell. Biol.23, 4257-4266.

21. Bhakat, K., Izumi, T., Yang, S., Hazra, T., and Mitra, S. (2003)EMBO J.,in press.

22. Izumi, T., Malecki, J., Chaudhry, M. A., Weinfeld, M., Hill, J.H., Lee, J. C., and Mitra, S. (1999)J. Mol. Biol. 287, 47-57.

23. Demple, B., and Harrison, L. (1994)Annu. ReV. Biochem. 63,915-948.

24. Masuda, Y., Bennett, R. A., and Demple, B. (1998)J. Biol. Chem.273, 30360-30365.

25. Masuda, Y., Bennett, R. A., and Demple, B. (1998)J. Biol. Chem.273, 30352-30359.

26. Beernink, P. T., Segelke, B. W., Hadi, M. Z., Erzberger, J. P.,Wilson, D. M., 3rd, and Rupp, B. (2001)J. Mol. Biol. 307, 1023-1034.

27. Gorman, M. A., Morera, S., Rothwell, D. G., de La Fortelle, E.,Mol, C. D., Tainer, J. A., Hickson, I. D., and Freemont, P. S.(1997)EMBO J. 16, 6548-6558.

28. Mol, C. D., Izumi, T., Mitra, S., and Tainer, J. A. (2000)Nature403, 451-456.

29. Schein, C. H., Oezgun, N., Izumi, T., and Braun, W. (2002)BMCBioinformatics 3, 37.

30. Izumi, T., and Mitra, S. (1998)Carcinogenesis 19, 525-527.31. Schein, C. H. (1989)Bio/Technology 7, 1141-1149.32. Schein, C. H. (1999)Protein Expression, Soluble, J. Wiley and

Sons, New York.33. Ramana, C. V., Boldogh, I., Izumi, T., Mitra, S. (1998)Proc. Natl.

Acad. Sci. U.S.A.95, 5061-5066.34. Hider, R., Drake, A., and Tamiya, N. (1988)Biopolymers 27, 113-

122.35. Wilson III, D. M., Takeshita1, M., Grollman, A. P., and Demple,

B. (1995)J. Biol. Chem. 270, 16002-16007.36. Wilson, D. M., 3rd, Takeshita, M., and Demple, B. (1997)Nucleic

Acids Res 25, 933-939.37. Wilson, S. H., and Kunkel, T. A. (2000)Nat. Struct. Biol. 7, 176-

178.38. Mol, C. D., Hosfield, D. J., and Tainer, J. A. (2000)Mutat. Res.

460, 211-229.39. Rothwell, D. G., and Hickson, I. D. (1996)Nucleic Acids Res.

24, 4217-4221.40. Tsujishita, Y., Guo, S., Stolz, L., York, J., and Hurley, J. (2001)

Cell 105, 379-389.

BI0346190

DNA Binding Specificity of Human APE1 Biochemistry, Vol. 43, No. 3, 2004689